Version 3.7.2

* Fix annoying rust-analyzer error with a cfg

* Refactor literal info into a struct

* Add field hide a literal from the GUI but still use for diffing

* Add config option to limit string literal diffing

* Fix deprecation warning `float_literal_f32_fallback`

* Hide non-preferred string encodings from the GUI

* Fix new cargo check warning

* Clippy again

* Rename "None" -> "Auto"

.github/workflows/build.yaml

@@ -62,7 +62,7 @@ jobs:
     strategy:
       matrix:
         checks:
-          - advisories
+          # - advisories
           - bans licenses sources
     # Prevent new advisories from failing CI
     continue-on-error: ${{ matrix.checks == 'advisories' }}

Cargo.lock

@@ -598,9 +598,9 @@ checksum = "8f1fe948ff07f4bd06c30984e69f5b4899c516a3ef74f34df92a2df2ab535495"
 
 [[package]]
 name = "bytes"
-version = "1.11.0"
+version = "1.11.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b35204fbdc0b3f4446b89fc1ac2cf84a8a68971995d0bf2e925ec7cd960f9cb3"
+checksum = "1e748733b7cbc798e1434b6ac524f0c1ff2ab456fe201501e6497c8417a4fc33"
 
 [[package]]
 name = "calloop"
@@ -3157,9 +3157,9 @@ dependencies = [
 
 [[package]]
 name = "num-conv"
-version = "0.1.0"
+version = "0.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "51d515d32fb182ee37cda2ccdcb92950d6a3c2893aa280e540671c2cd0f3b1d9"
+checksum = "cf97ec579c3c42f953ef76dbf8d55ac91fb219dde70e49aa4a6b7d74e9919050"
 
 [[package]]
 name = "num-traits"
@@ -3491,7 +3491,7 @@ dependencies = [
 
 [[package]]
 name = "objdiff-cli"
-version = "3.6.0"
+version = "3.7.2"
 dependencies = [
  "anyhow",
  "argp",
@@ -3514,7 +3514,7 @@ dependencies = [
 
 [[package]]
 name = "objdiff-core"
-version = "3.6.0"
+version = "3.7.2"
 dependencies = [
  "anyhow",
  "arm-attr",
@@ -3570,7 +3570,7 @@ dependencies = [
 
 [[package]]
 name = "objdiff-gui"
-version = "3.6.0"
+version = "3.7.2"
 dependencies = [
  "anyhow",
  "argp",
@@ -3608,7 +3608,7 @@ dependencies = [
 
 [[package]]
 name = "objdiff-wasm"
-version = "3.6.0"
+version = "3.7.2"
 dependencies = [
  "log",
  "objdiff-core",
@@ -5327,9 +5327,9 @@ dependencies = [
 
 [[package]]
 name = "time"
-version = "0.3.44"
+version = "0.3.47"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "91e7d9e3bb61134e77bde20dd4825b97c010155709965fedf0f49bb138e52a9d"
+checksum = "743bd48c283afc0388f9b8827b976905fb217ad9e647fae3a379a9283c4def2c"
 dependencies = [
  "deranged",
  "itoa",
@@ -5337,22 +5337,22 @@ dependencies = [
  "num-conv",
  "num_threads",
  "powerfmt",
- "serde",
+ "serde_core",
  "time-core",
  "time-macros",
 ]
 
 [[package]]
 name = "time-core"
-version = "0.1.6"
+version = "0.1.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "40868e7c1d2f0b8d73e4a8c7f0ff63af4f6d19be117e90bd73eb1d62cf831c6b"
+checksum = "7694e1cfe791f8d31026952abf09c69ca6f6fa4e1a1229e18988f06a04a12dca"
 
 [[package]]
 name = "time-macros"
-version = "0.2.24"
+version = "0.2.27"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "30cfb0125f12d9c277f35663a0a33f8c30190f4e4574868a330595412d34ebf3"
+checksum = "2e70e4c5a0e0a8a4823ad65dfe1a6930e4f4d756dcd9dd7939022b5e8c501215"
 dependencies = [
  "num-conv",
  "time-core",

Cargo.toml

@@ -14,7 +14,7 @@ default-members = [
 resolver = "3"
 
 [workspace.package]
-version = "3.6.0"
+version = "3.7.2"
 authors = ["Luke Street <luke@street.dev>"]
 edition = "2024"
 license = "MIT OR Apache-2.0"

objdiff-cli/src/cmd/diff.rs

@@ -198,8 +198,8 @@ fn run_oneshot(
         .transpose()?;
     let result =
         diff::diff_objs(target.as_ref(), base.as_ref(), None, &diff_config, &mapping_config)?;
-    let left = target.as_ref().and_then(|o| result.left.as_ref().map(|d| (o, d)));
-    let right = base.as_ref().and_then(|o| result.right.as_ref().map(|d| (o, d)));
+    let left = target.as_ref().zip(result.left.as_ref());
+    let right = base.as_ref().zip(result.right.as_ref());
     let diff_result = DiffResult::new(left, right, &diff_config)?;
     write_output(&diff_result, Some(output), output_format)?;
     Ok(())

objdiff-core/config-schema.json

@@ -117,6 +117,51 @@
       "name": "Combine text sections",
       "description": "Combines all text sections into one."
     },
+    {
+      "id": "preferredStringEncoding",
+      "type": "choice",
+      "default": "auto",
+      "name": "Preferred string encoding",
+      "description": "Which encoding to use when diffing string literals used in functions.",
+      "items": [
+        {
+          "value": "auto",
+          "name": "Auto"
+        },
+        {
+          "value": "ascii",
+          "name": "ASCII"
+        },
+        {
+          "value": "utf_8",
+          "name": "UTF-8"
+        },
+        {
+          "value": "shift_jis",
+          "name": "Shift JIS"
+        },
+        {
+          "value": "windows_1252",
+          "name": "Windows-1252"
+        },
+        {
+          "value": "euc_jp",
+          "name": "EUC-JP"
+        },
+        {
+          "value": "big5",
+          "name": "Big5"
+        },
+        {
+          "value": "utf_16be",
+          "name": "UTF-16BE"
+        },
+        {
+          "value": "utf_16le",
+          "name": "UTF-16LE"
+        }
+      ]
+    },
     {
       "id": "arm.archVersion",
       "type": "choice",
@@ -338,6 +383,7 @@
       "id": "general",
       "name": "General",
       "properties": [
+        "preferredStringEncoding",
         "functionRelocDiffs",
         "demangler",
         "showSymbolSizes",

objdiff-core/protos/diff.proto

@@ -39,7 +39,8 @@ message DiffSymbol {
   optional string demangled_name = 2;
   uint64 address = 3;
   uint64 size = 4;
-  uint32 flags = 5;
+  DiffSymbolFlags flags = 5;
+  DiffSymbolKind kind = 10;
   // The symbol index in the _other_ object that this symbol was diffed against
   optional uint32 target_symbol = 6;
   optional float match_percent = 7;
@@ -49,14 +50,22 @@ message DiffSymbol {
   repeated DiffDataSegment data_diff = 9;
 }
 
-// Symbol visibility flags (bitmask)
-enum DiffSymbolFlag {
-  SYMBOL_NONE = 0;
-  SYMBOL_GLOBAL = 1;
-  SYMBOL_LOCAL = 2;
-  SYMBOL_WEAK = 4;
-  SYMBOL_COMMON = 8;
-  SYMBOL_HIDDEN = 16;
+enum DiffSymbolKind {
+  SYMBOL_UNKNOWN = 0;
+  SYMBOL_FUNCTION = 1;
+  SYMBOL_OBJECT = 2;
+  SYMBOL_SECTION = 3;
+}
+
+// Symbol flags
+message DiffSymbolFlags {
+  bool global = 1;
+  bool local = 2;
+  bool weak = 3;
+  bool common = 4;
+  bool hidden = 5;
+  bool ignored = 6;
+  bool size_inferred = 7;
 }
 
 // A single instruction diff row

objdiff-core/src/arch/arm.rs

@@ -162,7 +162,7 @@ impl ArchArm {
 }
 
 impl Arch for ArchArm {
-    fn post_init(&mut self, sections: &[Section], symbols: &[Symbol]) {
+    fn post_init(&mut self, sections: &[Section], symbols: &[Symbol], _symbol_indices: &[usize]) {
         self.disasm_modes = Self::get_mapping_symbols(sections, symbols);
     }
 
@@ -224,7 +224,9 @@ impl Arch for ArchArm {
             }
 
             // Check how many bytes we can/should read
-            let num_code_bytes = if data.len() >= 4 {
+            let bytes_until_next_mapping =
+                next_mapping.map(|m| (m.address - address) as usize).unwrap_or(usize::MAX);
+            let num_code_bytes = if data.len() >= 4 && bytes_until_next_mapping >= 4 {
                 if mode == unarm::ParseMode::Data && address & 3 != 0 {
                     // 32-bit .word value should be aligned on a 4-byte boundary, otherwise use .hword
                     2
@@ -366,6 +368,22 @@ impl Arch for ArchArm {
                             (imm22 << 1) << 9 >> 9
                         }
 
+                        // Thumb unconditional branch (B, 11-bit offset)
+                        elf::R_ARM_THM_PC11 => {
+                            let data = section_data[address..address + 2].try_into()?;
+                            let insn = self.endianness.read_u16_bytes(data) as i32;
+                            let imm11 = insn & 0x7ff;
+                            (imm11 << 1) << 20 >> 20
+                        }
+
+                        // Thumb conditional branch (B<cond>, 8-bit offset)
+                        elf::R_ARM_THM_PC9 => {
+                            let data = section_data[address..address + 2].try_into()?;
+                            let insn = self.endianness.read_u16_bytes(data) as i32;
+                            let imm8 = insn & 0xff;
+                            (imm8 << 1) << 23 >> 23
+                        }
+
                         // Data
                         elf::R_ARM_ABS32 => {
                             let data = section_data[address..address + 4].try_into()?;
@@ -399,6 +417,8 @@ impl Arch for ArchArm {
                 elf::R_ARM_PC24 => Some("R_ARM_PC24"),
                 elf::R_ARM_XPC25 => Some("R_ARM_XPC25"),
                 elf::R_ARM_CALL => Some("R_ARM_CALL"),
+                elf::R_ARM_THM_PC11 => Some("R_ARM_THM_PC11"),
+                elf::R_ARM_THM_PC9 => Some("R_ARM_THM_PC9"),
                 _ => None,
             },
             _ => None,
@@ -418,6 +438,8 @@ impl Arch for ArchArm {
                 elf::R_ARM_PC24 => 4,
                 elf::R_ARM_XPC25 => 4,
                 elf::R_ARM_CALL => 4,
+                elf::R_ARM_THM_PC11 => 2,
+                elf::R_ARM_THM_PC9 => 2,
                 _ => 1,
             },
             _ => 1,
@@ -544,7 +566,9 @@ impl unarm::FormatIns for ArgsFormatter<'_> {
                 | RelocationFlags::Elf(elf::R_ARM_THM_PC22)
                 | RelocationFlags::Elf(elf::R_ARM_PC24)
                 | RelocationFlags::Elf(elf::R_ARM_XPC25)
-                | RelocationFlags::Elf(elf::R_ARM_CALL) => {
+                | RelocationFlags::Elf(elf::R_ARM_CALL)
+                | RelocationFlags::Elf(elf::R_ARM_THM_PC11)
+                | RelocationFlags::Elf(elf::R_ARM_THM_PC9) => {
                     return self.write(InstructionPart::reloc());
                 }
                 _ => {}

objdiff-core/src/arch/mod.rs

@@ -15,7 +15,7 @@ use object::Endian as _;
 
 use crate::{
     diff::{
-        DiffObjConfig, DiffSide,
+        ConfigEnum, DiffObjConfig, DiffSide, PreferredStringEncoding,
         display::{ContextItem, HoverItem, InstructionPart},
     },
     obj::{
@@ -42,15 +42,42 @@ pub mod x86;
 pub const OPCODE_INVALID: u16 = u16::MAX;
 pub const OPCODE_DATA: u16 = u16::MAX - 1;
 
-const SUPPORTED_ENCODINGS: [(&encoding_rs::Encoding, &str); 7] = [
+const SUPPORTED_ENCODINGS_WITH_NULL_TERM: [(&encoding_rs::Encoding, &str); 5] = [
     (encoding_rs::UTF_8, "UTF-8"),
     (encoding_rs::SHIFT_JIS, "Shift JIS"),
-    (encoding_rs::UTF_16BE, "UTF-16BE"),
-    (encoding_rs::UTF_16LE, "UTF-16LE"),
     (encoding_rs::WINDOWS_1252, "Windows-1252"),
     (encoding_rs::EUC_JP, "EUC-JP"),
     (encoding_rs::BIG5, "Big5"),
 ];
+const SUPPORTED_ENCODINGS_NO_NULL_TERM: [(&encoding_rs::Encoding, &str); 2] =
+    [(encoding_rs::UTF_16BE, "UTF-16BE"), (encoding_rs::UTF_16LE, "UTF-16LE")];
+
+#[derive(Debug, Clone, Default, PartialEq)]
+pub struct LiteralInfo {
+    pub literal: String,
+    pub label_override: Option<String>,
+    pub copy_string: Option<String>,
+    pub hidden: bool, // Only used when the user hasn't set a preferred string encoding
+    pub is_string: bool,
+}
+
+impl LiteralInfo {
+    pub fn hidden(&self, diff_config: Option<&DiffObjConfig>) -> bool {
+        let Some(diff_config) = diff_config else {
+            return self.hidden;
+        };
+        if !self.is_string {
+            return self.hidden;
+        }
+        if diff_config.preferred_string_encoding == PreferredStringEncoding::Auto {
+            return self.hidden;
+        }
+        let Some(ref label) = self.label_override else {
+            return self.hidden;
+        };
+        *label != diff_config.preferred_string_encoding.name()
+    }
+}
 
 /// Represents the type of data associated with an instruction
 #[derive(PartialEq)]
@@ -83,18 +110,14 @@ impl fmt::Display for DataType {
 impl DataType {
     pub fn display_labels(&self, endian: object::Endianness, bytes: &[u8]) -> Vec<String> {
         let mut strs = Vec::new();
-        for (literal, label_override, _) in self.display_literals(endian, bytes) {
-            let label = label_override.unwrap_or_else(|| self.to_string());
-            strs.push(format!("{label}: {literal:?}"))
+        for lit_info in self.display_literals(endian, bytes) {
+            let label = lit_info.label_override.unwrap_or_else(|| self.to_string());
+            strs.push(format!("{}: {:?}", label, lit_info.literal))
         }
         strs
     }
 
-    pub fn display_literals(
-        &self,
-        endian: object::Endianness,
-        bytes: &[u8],
-    ) -> Vec<(String, Option<String>, Option<String>)> {
+    pub fn display_literals(&self, endian: object::Endianness, bytes: &[u8]) -> Vec<LiteralInfo> {
         let mut strs = Vec::new();
         if self.required_len().is_some_and(|l| bytes.len() < l) {
             log::warn!(
@@ -118,80 +141,106 @@ impl DataType {
         match self {
             DataType::Int8 => {
                 let i = i8::from_ne_bytes(bytes.try_into().unwrap());
-                strs.push((format!("{i:#x}"), None, None));
+                strs.push(LiteralInfo { literal: format!("{i:#x}"), ..Default::default() });
 
                 if i < 0 {
-                    strs.push((format!("{:#x}", ReallySigned(i)), None, None));
+                    strs.push(LiteralInfo {
+                        literal: format!("{:#x}", ReallySigned(i)),
+                        ..Default::default()
+                    });
                 }
             }
             DataType::Int16 => {
                 let i = endian.read_i16_bytes(bytes.try_into().unwrap());
-                strs.push((format!("{i:#x}"), None, None));
+                strs.push(LiteralInfo { literal: format!("{i:#x}"), ..Default::default() });
 
                 if i < 0 {
-                    strs.push((format!("{:#x}", ReallySigned(i)), None, None));
+                    strs.push(LiteralInfo {
+                        literal: format!("{:#x}", ReallySigned(i)),
+                        ..Default::default()
+                    });
                 }
             }
             DataType::Int32 => {
                 let i = endian.read_i32_bytes(bytes.try_into().unwrap());
-                strs.push((format!("{i:#x}"), None, None));
+                strs.push(LiteralInfo { literal: format!("{i:#x}"), ..Default::default() });
 
                 if i < 0 {
-                    strs.push((format!("{:#x}", ReallySigned(i)), None, None));
+                    strs.push(LiteralInfo {
+                        literal: format!("{:#x}", ReallySigned(i)),
+                        ..Default::default()
+                    });
                 }
             }
             DataType::Int64 => {
                 let i = endian.read_i64_bytes(bytes.try_into().unwrap());
-                strs.push((format!("{i:#x}"), None, None));
+                strs.push(LiteralInfo { literal: format!("{i:#x}"), ..Default::default() });
 
                 if i < 0 {
-                    strs.push((format!("{:#x}", ReallySigned(i)), None, None));
+                    strs.push(LiteralInfo {
+                        literal: format!("{:#x}", ReallySigned(i)),
+                        ..Default::default()
+                    });
                 }
             }
             DataType::Float => {
                 let bytes: [u8; 4] = bytes.try_into().unwrap();
-                strs.push((
-                    format!("{:?}f", match endian {
+                strs.push(LiteralInfo {
+                    literal: format!("{:?}f", match endian {
                         object::Endianness::Little => f32::from_le_bytes(bytes),
                         object::Endianness::Big => f32::from_be_bytes(bytes),
                     }),
-                    None,
-                    None,
-                ));
+                    ..Default::default()
+                });
             }
             DataType::Double => {
                 let bytes: [u8; 8] = bytes.try_into().unwrap();
-                strs.push((
-                    format!("{:?}", match endian {
+                strs.push(LiteralInfo {
+                    literal: format!("{:?}", match endian {
                         object::Endianness::Little => f64::from_le_bytes(bytes),
                         object::Endianness::Big => f64::from_be_bytes(bytes),
                     }),
-                    None,
-                    None,
-                ));
+                    ..Default::default()
+                });
             }
             DataType::Bytes => {
-                strs.push((format!("{bytes:#?}"), None, None));
+                strs.push(LiteralInfo { literal: format!("{bytes:#?}"), ..Default::default() });
             }
             DataType::String => {
-                // Special case to display (ASCII) as the label for ASCII-only strings.
-                let mut is_ascii = false;
-                if bytes.is_ascii()
-                    && let Ok(str) = str::from_utf8(bytes)
-                {
-                    let trimmed = str.trim_end_matches('\0');
-                    if !trimmed.is_empty() {
-                        let copy_string = escape_special_ascii_characters(trimmed);
-                        strs.push((trimmed.to_string(), Some("ASCII".into()), Some(copy_string)));
-                        is_ascii = true;
+                if let Some(nul_idx) = bytes.iter().position(|&c| c == b'\0') {
+                    let str_bytes = &bytes[..nul_idx];
+                    // Special case to display (ASCII) as the label for ASCII-only strings.
+                    let (cow, _, had_errors) = encoding_rs::UTF_8.decode(str_bytes);
+                    if !had_errors && cow.is_ascii() {
+                        let string = format!("{cow}");
+                        let copy_string = escape_special_ascii_characters(&string);
+                        strs.push(LiteralInfo {
+                            literal: string,
+                            label_override: Some("ASCII".into()),
+                            copy_string: Some(copy_string),
+                            hidden: false,
+                            is_string: true,
+                        });
+                    }
+                    for (encoding, encoding_name) in SUPPORTED_ENCODINGS_WITH_NULL_TERM {
+                        let (cow, _, had_errors) = encoding.decode(str_bytes);
+                        if !had_errors {
+                            let string = format!("{cow}");
+                            let copy_string = escape_special_ascii_characters(&string);
+                            // Avoid showing ASCII-only strings more than once if the encoding is ASCII-compatible.
+                            let hidden = encoding.is_ascii_compatible() && cow.is_ascii();
+                            strs.push(LiteralInfo {
+                                literal: string,
+                                label_override: Some(encoding_name.into()),
+                                copy_string: Some(copy_string),
+                                hidden,
+                                is_string: true,
+                            });
+                        }
                     }
                 }
 
-                for (encoding, encoding_name) in SUPPORTED_ENCODINGS {
-                    // Avoid showing ASCII-only strings more than once if the encoding is ASCII-compatible.
-                    if is_ascii && encoding.is_ascii_compatible() {
-                        continue;
-                    }
+                for (encoding, encoding_name) in SUPPORTED_ENCODINGS_NO_NULL_TERM {
                     let (cow, _, had_errors) = encoding.decode(bytes);
                     if had_errors {
                         continue;
@@ -199,11 +248,13 @@ impl DataType {
                     let trimmed = cow.trim_end_matches('\0');
                     if !trimmed.is_empty() {
                         let copy_string = escape_special_ascii_characters(trimmed);
-                        strs.push((
-                            trimmed.to_string(),
-                            Some(encoding_name.into()),
-                            Some(copy_string),
-                        ));
+                        strs.push(LiteralInfo {
+                            literal: trimmed.to_string(),
+                            label_override: Some(encoding_name.into()),
+                            copy_string: Some(copy_string),
+                            hidden: false,
+                            is_string: true,
+                        });
                     }
                 }
             }
@@ -340,7 +391,8 @@ impl dyn Arch {
 
 pub trait Arch: Any + Debug + Send + Sync {
     /// Finishes arch-specific initialization that must be done after sections have been combined.
-    fn post_init(&mut self, _sections: &[Section], _symbols: &[Symbol]) {}
+    fn post_init(&mut self, _sections: &[Section], _symbols: &[Symbol], _symbol_indices: &[usize]) {
+    }
 
     /// Generate a list of instructions references (offset, size, opcode) from the given code.
     ///

objdiff-core/src/arch/ppc/flow_analysis.rs

@@ -513,6 +513,8 @@ pub fn ppc_data_flow_analysis(
     )
 }
 
+// Note: This function only supports MWCC ASCII strings.
+// Other encodings and other compilers are currently not supported.
 fn get_string_data(obj: &Object, symbol_index: usize, offset: Simm) -> Option<&str> {
     if let Some(sym) = obj.symbols.get(symbol_index)
         && sym.name.starts_with("@stringBase")

objdiff-core/src/arch/ppc/mod.rs

@@ -238,9 +238,9 @@ impl Arch for ArchPpc {
                         addend: 0,
                     })),
                     |(_, reloc)| match reloc.target() {
-                        object::RelocationTarget::Symbol(index) => Ok(Some(RelocationOverride {
+                        object::RelocationTarget::Symbol(_) => Ok(Some(RelocationOverride {
                             target: RelocationOverrideTarget::Keep,
-                            addend: index.0 as u16 as i16 as i64,
+                            addend: 0,
                         })),
                         target => Err(anyhow!("Unsupported IMAGE_REL_PPC_PAIR target {target:?}")),
                     },
@@ -359,9 +359,12 @@ impl Arch for ArchPpc {
 
     fn guess_data_type(&self, resolved: ResolvedInstructionRef, bytes: &[u8]) -> Option<DataType> {
         if resolved.relocation.is_some_and(|r| {
-            r.symbol.name.starts_with("@stringBase") || r.symbol.name.starts_with("$SG")
+            r.symbol.name.starts_with("@stringBase")
+                || r.symbol.name.starts_with("@wstringBase")
+                || r.symbol.name.starts_with("$SG")
+                || r.symbol.demangled_name == Some("`string'".to_string())
         }) {
-            // Pooled string.
+            // Compiler-generated symbol name for a string or a pool of strings.
             return Some(DataType::String);
         }
         let opcode = powerpc::Opcode::from(resolved.ins_ref.opcode);
@@ -472,12 +475,27 @@ impl Arch for ArchPpc {
         }
         Ok(next_address.saturating_sub(symbol.address))
     }
+
+    fn post_init(&mut self, _sections: &[Section], _symbols: &[Symbol], symbol_indices: &[usize]) {
+        // Change the indices used as keys from the original symbol indices to the new symbol array indices
+        self.extab = Self::convert_extab_map_indices(self, symbol_indices);
+    }
 }
 
 impl ArchPpc {
     pub fn extab_for_symbol(&self, symbol_index: usize) -> Option<&ExceptionInfo> {
         self.extab.as_ref()?.get(&symbol_index)
     }
+
+    pub fn convert_extab_map_indices(
+        &self,
+        symbol_indices: &[usize],
+    ) -> Option<BTreeMap<usize, ExceptionInfo>> {
+        let new_map: BTreeMap<usize, ExceptionInfo> =
+            self.extab.as_ref()?.iter().map(|e| (symbol_indices[*e.0 + 1], e.1.clone())).collect();
+
+        Some(new_map)
+    }
 }
 
 fn zero_reloc(code: u32, reloc: &Relocation) -> u32 {

objdiff-core/src/bindings/diff.rs

@@ -43,9 +43,6 @@ impl DiffObject {
         let mut symbols = Vec::with_capacity(obj.symbols.len());
         for (symbol_idx, symbol) in obj.symbols.iter().enumerate() {
             let symbol_diff = &diff.symbols[symbol_idx];
-            if symbol.size == 0 || symbol.flags.contains(SymbolFlag::Ignored) {
-                continue;
-            }
             symbols.push(DiffSymbol::new(obj, symbol_idx, symbol, symbol_diff, diff_config)?);
         }
 
@@ -112,6 +109,7 @@ impl DiffSymbol {
             address: symbol.address,
             size: symbol.size,
             flags: symbol_flags(&symbol.flags),
+            kind: DiffSymbolKind::from(symbol.kind) as i32,
             // Diff information
             target_symbol: symbol_diff.target_symbol.map(|i| i as u32),
             match_percent: symbol_diff.match_percent,
@@ -121,24 +119,27 @@ impl DiffSymbol {
     }
 }
 
-fn symbol_flags(flags: &obj::SymbolFlagSet) -> u32 {
-    let mut result = 0u32;
-    if flags.contains(SymbolFlag::Global) {
-        result |= DiffSymbolFlag::SymbolGlobal as u32;
+impl From<obj::SymbolKind> for DiffSymbolKind {
+    fn from(value: obj::SymbolKind) -> Self {
+        match value {
+            obj::SymbolKind::Unknown => DiffSymbolKind::SymbolUnknown,
+            obj::SymbolKind::Function => DiffSymbolKind::SymbolFunction,
+            obj::SymbolKind::Object => DiffSymbolKind::SymbolObject,
+            obj::SymbolKind::Section => DiffSymbolKind::SymbolSection,
+        }
     }
-    if flags.contains(SymbolFlag::Local) {
-        result |= DiffSymbolFlag::SymbolLocal as u32;
-    }
-    if flags.contains(SymbolFlag::Weak) {
-        result |= DiffSymbolFlag::SymbolWeak as u32;
-    }
-    if flags.contains(SymbolFlag::Common) {
-        result |= DiffSymbolFlag::SymbolCommon as u32;
-    }
-    if flags.contains(SymbolFlag::Hidden) {
-        result |= DiffSymbolFlag::SymbolHidden as u32;
-    }
-    result
+}
+
+fn symbol_flags(flags: &obj::SymbolFlagSet) -> Option<DiffSymbolFlags> {
+    Some(DiffSymbolFlags {
+        global: flags.contains(SymbolFlag::Global),
+        local: flags.contains(SymbolFlag::Local),
+        weak: flags.contains(SymbolFlag::Weak),
+        common: flags.contains(SymbolFlag::Common),
+        hidden: flags.contains(SymbolFlag::Hidden),
+        ignored: flags.contains(SymbolFlag::Ignored),
+        size_inferred: flags.contains(SymbolFlag::SizeInferred),
+    })
 }
 
 impl DiffInstructionRow {

objdiff-core/src/diff/code.rs

@@ -9,8 +9,8 @@ use anyhow::{Context, Result, anyhow, ensure};
 
 use super::{
     DiffObjConfig, FunctionRelocDiffs, InstructionArgDiffIndex, InstructionBranchFrom,
-    InstructionBranchTo, InstructionDiffKind, InstructionDiffRow, SymbolDiff,
-    display::display_ins_data_literals,
+    InstructionBranchTo, InstructionDiffKind, InstructionDiffRow, PreferredStringEncoding,
+    SymbolDiff, display::display_ins_data_literals,
 };
 use crate::obj::{
     InstructionArg, InstructionArgValue, InstructionRef, Object, ResolvedInstructionRef,
@@ -296,6 +296,26 @@ pub(crate) fn section_name_eq(
     })
 }
 
+fn ins_data_literals_eq(
+    left_obj: &Object,
+    right_obj: &Object,
+    left_ins: ResolvedInstructionRef,
+    right_ins: ResolvedInstructionRef,
+    diff_config: &DiffObjConfig,
+) -> bool {
+    let mut left_literals = display_ins_data_literals(left_obj, left_ins);
+    let mut right_literals = display_ins_data_literals(right_obj, right_ins);
+    if left_literals == right_literals {
+        return true;
+    }
+    if diff_config.preferred_string_encoding == PreferredStringEncoding::Auto {
+        return left_literals == right_literals;
+    }
+    left_literals.retain(|lit_info| !lit_info.hidden(Some(diff_config)));
+    right_literals.retain(|lit_info| !lit_info.hidden(Some(diff_config)));
+    left_literals == right_literals
+}
+
 fn reloc_eq(
     left_obj: &Object,
     right_obj: &Object,
@@ -330,8 +350,7 @@ fn reloc_eq(
                 && (diff_config.function_reloc_diffs == FunctionRelocDiffs::NameAddress
                     || left_reloc.symbol.kind != SymbolKind::Object
                     || right_reloc.symbol.size == 0 // Likely a pool symbol like ...data, don't treat this as a diff
-                    || display_ins_data_literals(left_obj, left_ins)
-                        == display_ins_data_literals(right_obj, right_ins))
+                    || ins_data_literals_eq(left_obj, right_obj, left_ins, right_ins, diff_config))
         }
         (Some(_), None) | (None, Some(_)) | (None, None) => symbol_name_addend_matches,
     }

objdiff-core/src/diff/display.rs

@@ -12,6 +12,7 @@ use itertools::Itertools;
 use regex::Regex;
 
 use crate::{
+    arch::LiteralInfo,
     diff::{
         DataDiffKind, DataDiffRow, DiffObjConfig, InstructionDiffKind, InstructionDiffRow,
         ObjectDiff, SymbolDiff, data::resolve_relocation,
@@ -494,15 +495,21 @@ pub fn relocation_context(
     obj: &Object,
     reloc: ResolvedRelocation,
     ins: Option<ResolvedInstructionRef>,
+    diff_config: Option<&DiffObjConfig>,
 ) -> Vec<ContextItem> {
     let mut out = Vec::new();
     out.append(&mut symbol_context(obj, reloc.relocation.target_symbol));
     if let Some(ins) = ins {
-        let literals = display_ins_data_literals(obj, ins);
+        let mut literals = display_ins_data_literals(obj, ins);
+        literals.retain(|lit_info| !lit_info.hidden(diff_config));
         if !literals.is_empty() {
             out.push(ContextItem::Separator);
-            for (literal, label_override, copy_string) in literals {
-                out.push(ContextItem::Copy { value: literal, label: label_override, copy_string });
+            for lit_info in literals {
+                out.push(ContextItem::Copy {
+                    value: lit_info.literal,
+                    label: lit_info.label_override,
+                    copy_string: lit_info.copy_string,
+                });
             }
         }
     }
@@ -555,7 +562,7 @@ pub fn data_row_context(obj: &Object, diff_row: &DataDiffRow) -> Vec<ContextItem
         prev_reloc = Some(reloc);
 
         let reloc = resolve_relocation(&obj.symbols, reloc);
-        out.append(&mut relocation_context(obj, reloc, None));
+        out.append(&mut relocation_context(obj, reloc, None, None));
     }
     out
 }
@@ -592,6 +599,7 @@ pub fn instruction_context(
     obj: &Object,
     resolved: ResolvedInstructionRef,
     ins: &ParsedInstruction,
+    diff_config: &DiffObjConfig,
 ) -> Vec<ContextItem> {
     let mut out = Vec::new();
     let mut hex_string = String::new();
@@ -636,7 +644,7 @@ pub fn instruction_context(
     }
     if let Some(reloc) = resolved.relocation {
         out.push(ContextItem::Separator);
-        out.append(&mut relocation_context(obj, reloc, Some(resolved)));
+        out.append(&mut relocation_context(obj, reloc, Some(resolved), Some(diff_config)));
     }
     out
 }
@@ -645,6 +653,7 @@ pub fn instruction_hover(
     obj: &Object,
     resolved: ResolvedInstructionRef,
     ins: &ParsedInstruction,
+    diff_config: &DiffObjConfig,
 ) -> Vec<HoverItem> {
     let mut out = Vec::new();
     out.push(HoverItem::Text {
@@ -687,13 +696,14 @@ pub fn instruction_hover(
         out.append(&mut relocation_hover(obj, reloc, None));
         let bytes = obj.symbol_data(reloc.relocation.target_symbol).unwrap_or(&[]);
         if let Some(ty) = obj.arch.guess_data_type(resolved, bytes) {
-            let literals = display_ins_data_literals(obj, resolved);
+            let mut literals = display_ins_data_literals(obj, resolved);
+            literals.retain(|lit_info| !lit_info.hidden(Some(diff_config)));
             if !literals.is_empty() {
                 out.push(HoverItem::Separator);
-                for (literal, label_override, _) in literals {
+                for lit_info in literals {
                     out.push(HoverItem::Text {
-                        label: label_override.unwrap_or_else(|| ty.to_string()),
-                        value: format!("{literal:?}"),
+                        label: lit_info.label_override.unwrap_or_else(|| ty.to_string()),
+                        value: format!("{:?}", lit_info.literal),
                         color: HoverItemColor::Normal,
                     });
                 }
@@ -880,7 +890,7 @@ pub fn display_ins_data_labels(obj: &Object, resolved: ResolvedInstructionRef) -
 pub fn display_ins_data_literals(
     obj: &Object,
     resolved: ResolvedInstructionRef,
-) -> Vec<(String, Option<String>, Option<String>)> {
+) -> Vec<LiteralInfo> {
     let Some(reloc) = resolved.relocation else {
         return Vec::new();
     };

objdiff-core/src/jobs/create_scratch.rs

@@ -5,6 +5,7 @@ use typed_path::{Utf8PlatformPathBuf, Utf8UnixPathBuf};
 
 use crate::{
     build::{BuildConfig, BuildStatus, run_make},
+    diff::Demangler,
     jobs::{Job, JobContext, JobResult, JobState, start_job, update_status},
 };
 
@@ -21,6 +22,7 @@ pub struct CreateScratchConfig {
     pub function_name: String,
     pub target_obj: Utf8PlatformPathBuf,
     pub preset_id: Option<u32>,
+    pub demangler: Demangler,
 }
 
 #[derive(Default, Debug, Clone)]
@@ -74,7 +76,14 @@ fn run_create_scratch(
         .text("diff_label", config.function_name.clone())
         .text("diff_flags", diff_flags)
         .text("context", context.unwrap_or_default())
-        .text("source_code", "// Move related code from Context tab to here");
+        .text("source_code", "// Move related code from Context tab to here")
+        .text(
+            "name",
+            config
+                .demangler
+                .demangle(&config.function_name)
+                .unwrap_or_else(|| config.function_name.clone()),
+        );
     if let Some(preset) = config.preset_id {
         form = form.text("preset", preset.to_string());
     }

objdiff-core/src/jobs/objdiff.rs

@@ -175,8 +175,8 @@ fn run_build(
     Ok(Box::new(ObjDiffResult {
         first_status,
         second_status,
-        first_obj: first_obj.and_then(|o| result.left.map(|d| (o, d))),
-        second_obj: second_obj.and_then(|o| result.right.map(|d| (o, d))),
+        first_obj: first_obj.zip(result.left),
+        second_obj: second_obj.zip(result.right),
         time,
     }))
 }

objdiff-core/src/obj/read.rs

@@ -40,6 +40,7 @@ fn map_section_kind(section: &object::Section) -> SectionKind {
 /// e.g. symbol$1234 and symbol$2345 will both be replaced with symbol$0000 internally.
 fn get_normalized_symbol_name(name: &str) -> Option<String> {
     const DUMMY_UNIQUE_ID: &str = "0000";
+    const DUMMY_UNIQUE_MSVC_ID: &str = "00000000";
     if let Some((prefix, suffix)) = name.split_once("@class$")
         && let Some(idx) = suffix.chars().position(|c| !c.is_numeric())
         && idx > 0
@@ -59,6 +60,26 @@ fn get_normalized_symbol_name(name: &str) -> Option<String> {
     {
         // Match GCC symbol.1234 against symbol.2345
         Some(format!("{prefix}.{DUMMY_UNIQUE_ID}"))
+    } else if name.starts_with('?') {
+        // Match MSVC anonymous class symbol names, ignoring the unique ID.
+        // e.g. ?CheckContextOr@?A0x24773155@@YA_NPBVDataArray@@@Z
+        // and: ?CheckContextOr@?A0xddf6240c@@YA_NPBVDataArray@@@Z
+        let mut name_str = String::from(name);
+        let anon_indices: Vec<usize> = name_str.match_indices("?A0x").map(|(idx, _)| idx).collect();
+        if !anon_indices.is_empty() {
+            for idx in anon_indices {
+                // the str sequence we're looking for is: ?A0xXXXXXXXX@@
+                if u32::from_str_radix(&name_str[idx + 4..idx + 12], 16).is_ok()
+                    && &name_str[idx + 12..idx + 14] == "@@"
+                {
+                    // if the two above checks passed, we're good to replace the hash
+                    name_str.replace_range(idx + 4..idx + 12, DUMMY_UNIQUE_MSVC_ID);
+                }
+            }
+            Some(name_str)
+        } else {
+            None
+        }
     } else {
         None
     }
@@ -874,8 +895,12 @@ fn combine_sections(
     let mut data_sections = BTreeMap::<String, Vec<usize>>::new();
     let mut text_sections = BTreeMap::<String, Vec<usize>>::new();
     for (i, section) in sections.iter().enumerate() {
-        let base_name =
-            if let Some(i) = section.name.rfind('$') { &section.name[..i] } else { &section.name };
+        let base_name = section
+            .name
+            .get(1..)
+            .and_then(|s| s.rfind(['$', '.']))
+            .and_then(|i| section.name.get(..i + 1))
+            .unwrap_or(&section.name);
         match section.kind {
             SectionKind::Data | SectionKind::Bss => {
                 data_sections.entry(base_name.to_string()).or_default().push(i);
@@ -1064,7 +1089,7 @@ pub fn parse(data: &[u8], config: &DiffObjConfig, diff_side: DiffSide) -> Result
         combine_sections(&mut sections, &mut symbols, config)?;
     }
     add_section_symbols(&sections, &mut symbols);
-    arch.post_init(&sections, &symbols);
+    arch.post_init(&sections, &symbols, &symbol_indices);
     let mut obj = Object {
         arch,
         endianness: obj_file.endianness(),

objdiff-core/tests/arch_arm.rs

@@ -57,6 +57,31 @@ fn combine_text_sections() {
     insta::assert_snapshot!(output);
 }
 
+#[test]
+#[cfg(feature = "arm")]
+fn thumb_short_data_mapping() {
+    // When a .2byte directive is used in Thumb code, the assembler emits
+    // $d/$t mapping symbols for a 2-byte data region. The disassembler must
+    // not read 4 bytes as a .word when the next mapping symbol limits the
+    // data region to 2 bytes.
+    let diff_config = diff::DiffObjConfig::default();
+    let obj = obj::read::parse(
+        include_object!("data/arm/code_1_vblank.o"),
+        &diff_config,
+        diff::DiffSide::Base,
+    )
+    .unwrap();
+    let symbol_idx = obj.symbols.iter().position(|s| s.name == "VBlankDMA_Level1").unwrap();
+    let diff = diff::code::no_diff_code(&obj, symbol_idx, &diff_config).unwrap();
+    let output = common::display_diff(&obj, &diff, symbol_idx, &diff_config);
+    // .2byte data followed by Thumb code must not be merged into a 4-byte .word
+    assert!(
+        !output.contains(".word"),
+        "2-byte data regions should not be decoded as 4-byte .word values.\n\
+         The disassembler must respect mapping symbol boundaries."
+    );
+}
+
 #[test]
 #[cfg(feature = "arm")]
 fn trim_trailing_hword() {

objdiff-core/tests/arch_ppc.rs

@@ -122,3 +122,90 @@ fn read_vmx128_coff() {
     let output = common::display_diff(&obj, &diff, symbol_idx, &diff_config);
     insta::assert_snapshot!(output);
 }
+
+#[test]
+#[cfg(feature = "ppc")]
+fn decode_sjis_pooled_strings() {
+    // Test multiple pooled Shift JIS strings separated by a single null byte between entries. (MWCC)
+    let diff_config = diff::DiffObjConfig { combine_data_sections: true, ..Default::default() };
+    let obj = obj::read::parse(
+        include_object!("data/ppc/m_Do_hostIO.o"),
+        &diff_config,
+        diff::DiffSide::Base,
+    )
+    .unwrap();
+    common::assert_literal_value(
+        &obj,
+        &diff_config,
+        "createChild__16mDoHIO_subRoot_cFPCcP13JORReflexible",
+        15,
+        "Shift JIS",
+        "危険：既に登録されているホストIOをふたたび登録しようとしています<%s>\n",
+    );
+    common::assert_literal_value(
+        &obj,
+        &diff_config,
+        "createChild__16mDoHIO_subRoot_cFPCcP13JORReflexible",
+        42,
+        "Shift JIS",
+        "ホストIOの空きエントリがありません。登録できませんでした。\n",
+    );
+}
+
+#[test]
+#[cfg(feature = "ppc")]
+fn decode_utf16_unpooled_strings() {
+    // Test unpooled UTF-16BE wide strings with null bytes at the start, end, and in the middle of the string. (MSVC)
+    let diff_config = diff::DiffObjConfig { combine_data_sections: true, ..Default::default() };
+    let obj = obj::read::parse(
+        include_object!("data/ppc/KinectSharePanel.obj"),
+        &diff_config,
+        diff::DiffSide::Base,
+    )
+    .unwrap();
+    common::assert_literal_value(
+        &obj,
+        &diff_config,
+        "?OnPostLink@KinectSharePanel@@AAA?AVDataNode@@PAVDataArray@@@Z",
+        84,
+        "UTF-16BE",
+        "Title Text",
+    );
+    common::assert_literal_value(
+        &obj,
+        &diff_config,
+        "?OnPostLink@KinectSharePanel@@AAA?AVDataNode@@PAVDataArray@@@Z",
+        120,
+        "UTF-16BE",
+        "http://www.dancecentral.com/content-assets/2012/06/2012E3LogoBox_tn.jpg",
+    );
+}
+
+#[test]
+#[cfg(feature = "ppc")]
+fn decode_ascii_strings_with_null_padding() {
+    // Test unpooled ASCII strings with more than one null byte at the end.
+    let diff_config = diff::DiffObjConfig { combine_data_sections: true, ..Default::default() };
+    let obj = obj::read::parse(
+        include_object!("data/ppc/vmx128.obj"),
+        &diff_config,
+        diff::DiffSide::Base,
+    )
+    .unwrap();
+    common::assert_literal_value(
+        &obj,
+        &diff_config,
+        "?PrintVector@@YAXPBDU__vector4@@@Z",
+        24,
+        "ASCII",
+        "%s: [%.2f, %.2f, %.2f, %.2f]\n",
+    );
+    common::assert_literal_value(
+        &obj,
+        &diff_config,
+        "?ReservedRegisterExample@@YAXXZ",
+        59,
+        "ASCII",
+        "Result from vr66",
+    );
+}

objdiff-core/tests/common.rs

@@ -50,3 +50,24 @@ macro_rules! include_object {
         include_bytes_align_as!(u64, $path)
     };
 }
+
+#[allow(dead_code)]
+pub fn assert_literal_value(
+    obj: &Object,
+    diff_config: &DiffObjConfig,
+    function_name: &str,
+    ins_row_idx: usize,
+    literal_label: &str,
+    literal_value: &str,
+) {
+    let symbol_idx = obj.symbols.iter().position(|s| s.name == function_name).unwrap();
+    let diff = objdiff_core::diff::code::no_diff_code(obj, symbol_idx, diff_config).unwrap();
+    let ins_ref = diff.instruction_rows[ins_row_idx].ins_ref.unwrap();
+    let resolved = obj.resolve_instruction_ref(symbol_idx, ins_ref).unwrap();
+    let literals = objdiff_core::diff::display::display_ins_data_literals(obj, resolved);
+    let target_literal =
+        literals.iter().find(|lit_info| lit_info.label_override == Some(literal_label.to_string()));
+    assert_ne!(target_literal, None);
+    let lit_info = target_literal.unwrap();
+    assert_eq!(lit_info.literal, literal_value);
+}